From clicks to conversations: How Claiva used an agentic approach on Amazon Bedrock to personalize every travel journey
Claiva, a London-based travel technology company, partnered with AWS Generative AI Innovation Center to transform their price notification system into an intelligent AI-powered travel companion using Amazon Bedrock. In just 6 weeks, they developed a new generative AI product that provides conversational recommendations and contextual guidance to travelers, accelerating their product roadmap by 6 months. In production, this new solution 27% booking intent increase for major UK travel platform.
The travel industry is experiencing a fundamental shift in how customers discover and book their next adventure. With over 1.4 billion international travelers expected in 2025, travel companies face the challenge of providing personalized experiences at scale. In this work we offer deep personalized guidance to travelers in addition to the price and availability recommendations offered by traditional online booking platforms so that travelers need to make confident decisions.
Claiva, a London-based travel technology provider, recognized this gap and partnered with the AWS Generative AI Innovation Center to revolutionize how online travel agencies engage with their customers.
In 6 weeks Claiva developed a new Generative AI product offering, accelerating their roadmap by 6 months ... Through our partnership with the AWS Generative AI Innovation Center, we've built a platform that bridges the critical gap between fluctuating prices, customer aspirations, and booking confidence." says Nic Costa, CEO at Claiva.
By leveraging Amazon Bedrock and other AWS services, Claiva have transformed their existing WhatsApp-based notification system PriceWatcher into an intelligent travel companion that understands customer preferences and provides contextual recommendations in real-time.
The Challenge: Beyond Basic Price Alerts
Claiva's initial tests used price tracking and notifications through an early MVP on WhatsApp, demonstrating the feasibility of real-time recommendations via chat.
However, the system was limited in its ability to:
- Provide detailed natural language context around price changes
- Maintain engaging long-running conversations
- Deliver truly personalized recommendations based on customer preferences
- Scale personalized interactions across a growing customer base
The AWS Generative AI Innovation Center worked with Claiva to design and develop a proof of concept that would demonstrate how generative AI could transform these basic notifications into intelligent conversations. The solution needed to combine real-time pricing data with natural language understanding to create a more engaging and helpful booking experience.
Solution
To transform price notifications into meaningful conversations, Claiva worked with the AWS Generative AI Innovation Center to architect a solution that uses Amazon Bedrock's foundation models alongside a suite of AWS services.
The solution implements an intelligent agent architecture that maintains context, manages conversation flow, and delivers personalized recommendations through structured data handling and flexible prompt engineering. The solution combines the power of Amazon Bedrock's foundation models with scalable serverless architecture on AWS shown in the attached image.
The solution consists of five main components; an Intelligent Conversation Engine, User Profile Generator, User Research, Package Recommendation Engine and Agent Builder UI.
Intelligent Conversation Engine
This component orchestrates a long-running conversation with the user.
- Amazon Bedrock foundation models serve as the core orchestration engine, managing multi-turn conversations and long-running interactions.
- Custom Lambda functions handle business logic and integrate with external services implementing the logic for generating user research and recommending packages.
- Structured data handling through Pydantic ensures type safety and robustness.
User Profile Generator
This component generates a distinct user profile based on their conversational interactions.
- Hybrid profile structure combining structured fields for reliable data storage.
- Unstructured notes capture nuanced preferences and contextual information.
Progressive profile enrichment through natural conversation.
User Research
This system generates unique user research questions for deeper user understanding.
- A dedicated AWS Lambda function generates a user research question and a variable number of potential answers, reducing friction for the user to provide their preferences.
- The question is customised based on what we have already collected in their user profile and other information from their dictated travel package preferences.
Package Recommendation Engine
This recommends packages to the user based on their user profile.
- External tool calls integrate real-time pricing data
- Contextual filtering based on user preferences and conversation history
- Dynamic re-ranking of recommendations based on user feedback
Agent Builder UI
The agent builder UI provides a lightweight streamlit-based demo interface for testing and prototyping new agent configurations before deploying to production. The below configuration demonstrates how the solution can be customised to interact with users via a range of different configuration instructions. Here, we can instruct the agent to always ask certain questions and influence when it will attempt to make a first package recommendation in the user interaction.
Conclusion
The travel industry continues to evolve, with personalization and real-time engagement becoming key differentiators for successful businesses. Through their collaboration with the AWS Generative AI Innovation Center, Claiva has demonstrated how Amazon Bedrock can transform traditional price tracking into intelligent, contextual conversations that drive better booking decisions.
The intelligent agent architecture on AWS has delivered tangible results for Claiva's clients including a 27% increase in customer booking intent for travel packages.
By combining generative AI capabilities with real-time pricing data, Claiva created a solution that doesn't just inform travelers about price changes – it helps them understand the context and make confident decisions about their travel plans. This shift from transactional alerts to conversational engagement represents the future of travel technology, where every interaction is meaningful.
Ready to transform your customer experience with generative AI? Visit Claiva's website to learn more about their travel solutions, or connect with the AWS Generative AI Innovation Center to explore how you can build intelligent, conversational applications for your use case with Amazon Bedrock.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article